Study on the Impact Mechanism of Enterprise Digital Transformation on Supply Chain Resilience

Abstract

This study examines how digital transformation enhances supply chain resilience among Chinese firms, with a focus on the underlying mechanisms and contextual conditions. Grounded in dynamic capabilities theory, we conceptualize supply chain resilience along two dimensions: proactive capability and reactive capability. Using data from A-share listed companies between 2007 and 2022, we construct firm-level resilience measures through entropy weighting. Digital transformation is measured by textual analysis of corporate annual reports, supplemented with policy documents and academic literature to enrich the keyword dictionary. Empirical results, validated through instrumental variable estimation, Heckman two-stage models, and multiple robustness checks, show that digital transformation significantly improves overall supply chain resilience, with a stronger effect on reactive capability. Further analysis identifies three mediating channels: improved information sharing across the supply chain, enhanced firm-level innovation, and reduced exposure to environmental uncertainty. Heterogeneity tests reveal that the positive impact of digital transformation is more pronounced in non-state-owned enterprises, high-tech firms, and firms in technology-intensive or labor-intensive industries. The effect is also stronger for firms operating under high environmental uncertainty or located in regions with lower levels of marketization. These findings offer practical guidance for managers and policymakers aiming to strengthen supply chains through digitalization, particularly in an era marked by growing global disruptions and sustainability challenges.

1. Introduction

As global supply chains face rising geopolitical tensions, accelerating technological decoupling, and more frequent systemic shocks, supply chain resilience is increasingly vital. It not only supports industrial stability but also underpins high-quality and sustainable economic development [1]. In China, the dual challenges of external technology blockades and internal structural upgrading have exposed critical “bottlenecks”, “breakpoints”, and capability gaps in supply systems, hindering the smooth functioning of the dual-circulation strategy and the transition toward high-quality, green growth [2]. While digital technologies such as big data, artificial intelligence, and blockchain are increasingly embedded in corporate operations, offering new pathways to enhance responsiveness and efficiency, a significant theoretical gap remains: how exactly does enterprise-level digital transformation translate into greater supply chain resilience through internal capability mechanisms? Existing research often treats digitalization as a black box or focuses predominantly on macro-level outcomes, thereby neglecting the micro-foundations that connect digital strategy to dynamic capabilities, which are essential to resilient and, by extension, sustainable supply networks [3].

This study directly addresses a central question: Can and how does enterprise digital transformation enhance supply chain resilience by reconfiguring key resources and optimizing supply chain dynamic capabilities? Grounded in the resource-based view and dynamic capabilities theory, we conceptualize supply chain resilience as jointly determined by two dimensions of dynamic capability: proactive capability (e.g., risk anticipation, strategic sensing) and reactive capability (e.g., rapid recovery, adaptive adjustment). Using a comprehensive sample of Chinese A-share listed firms from 2007 to 2022, we construct a composite resilience index via entropy weighting based on these two dimensions. Digital transformation is measured through textual analysis of annual reports using a validated dictionary of 76 digital-related keywords across five domains. We further test three specific mediating channels through which digital transformation operates: (1) enhanced information spillovers across supply chain partners, (2) improved firm-level innovation capacity, and (3) reduced exposure to environmental uncertainty. Additionally, we conduct heterogeneity analyses across industry characteristics, ownership type, marketization level, and environmental uncertainty regimes to identify boundary conditions.

The main marginal contributions of this paper are as follows: First, from a perspective standpoint, we shift the focus from macro or static assessments to a micro-level, dynamic capabilities-based framework, explicitly decomposing supply chain resilience into proactive and reactive sub-dimensions. This study aims to help address a gap in the literature regarding the internal capability mechanisms through which digital transformation operates. Second, from a theoretical perspective, we propose and empirically test a novel tripartite mediation model comprising information spillover, innovation enhancement, and environmental uncertainty reduction. This model helps unpack the “black box” linking digitalization to supply chain resilience. Third, in practical terms, we identify and empirically validate three specific mechanisms through which digitalization enhances resilience, offering firms actionable levers to implement sustainable operations. Combined with rigorous heterogeneity analyses, these findings provide an evidence base for designing policies that support green-digital transitions [4].

The following sections of this paper are organized as follows: Section 2 includes a comprehensive review of the literature; Section 3 presents theoretical analysis and research hypotheses; Section 4 outlines the research design, including model construction, variable design, and data sources; Section 5 consists of empirical results and analysis; and Section 6 introduces the conclusion, practical implications, and limitations.

2. Literature Review

2.1. Literature Review on Digital Transformation of Enterprises

The academic understanding of enterprise digital transformation has undergone a fundamental shift, evolving from an early “tool-centric” perspective to a “strategic eco-system” view. Early research often narrowly defined it as the application of digital technologies in specific business scenarios to optimize processes [5]. However, contemporary scholars argue that its essence is a multi-dimensional systematic change covering technology, organization, and management [6]. Digital transformation is not merely a technological upgrade but a fundamental reshaping of the organizational ecology, aiming to generate continuous innovation and thoroughly revolutionize the development momentum of enterprises [7]. To capture this strategic shift, measurement methodologies have also evolved from static indicators to dynamic tracking. While early studies primarily relied on survey-based quantification of IT investment or simple dummy variable assignment [8], recent literature increasingly favors the textual analysis of annual reports. By tracking the frequency of keywords, this method can more accurately capture the strategic intensity and depth of corporate transformation efforts [9].

In terms of economic consequences, digital transformation generates impacts that progressively expand from internal operations to external ecosystems. First, at the level of internal productivity, digital transformation improves Total Factor Productivity (TFP) by stimulating innovation vitality, optimizing human capital structure [7], and promoting the integration of advanced manufacturing with modern services [10]. Second, regarding the information environment, digitalization visualizes internal data, which significantly reduces information asymmetry between the enterprise and external stakeholders. This reduction in asymmetry enhances positive market expectations and optimizes capital structure, indicating that digital transformation effectively acts as a mechanism for information transmission [11]. Finally, extending to the supply chain network, these efficiencies and information gains translate into stronger systemic capabilities. Digitalization strengthens supply chain integration and link synergy, allowing for better capital deployment and risk resistance [12]. By breaking down data silos, it constructs a collaborative innovation ecosystem that boosts the overall performance and sustainability of the enterprise [13].

2.2. Literature Review on Supply Chain Resilience

The theoretical framework of supply chain resilience has evolved from a focus on static post-disruption recovery to a dynamic perspective of adaptive evolution. Early definitions primarily characterized resilience as the capacity of a system to absorb shocks and return to its initial equilibrium state after a disruption [14]. However, contemporary research grounded in dynamic capability theory argues that resilience extends beyond reactive restoration; it encompasses the comprehensive ability to anticipate, monitor, react, and learn, enabling the supply chain to evolve to a more robust state post-disruption [15,16]. This conceptual shift has directly driven the evolution of measurement methodologies from single outcome-based metrics to multi-dimensional evaluations. Scholars now assess resilience by combining performance indicators, such as recovery time and loss degree [17], with capability indicators like collaboration efficiency and agility, utilizing integrated evaluation systems to comprehensively capture this complexity [18].

Regarding the determinants of resilience, existing literature identifies two primary influencing mechanisms: the Resource-Based View (RBV) and the Dynamic Capability Perspective. From the resource perspective, resilience depends on the strategic buffering and visibility of assets. This involves not only the deployment of tangible redundant resources, such as safety stock, to absorb immediate shocks, but also the utilization of intangible digital infrastructure. Specifically, information management systems act as critical resources that facilitate real-time information sharing, thereby significantly enhancing the visibility and response speed required to manage disruptions [19]. In contrast, the capability perspective emphasizes that resources must be mobilized by dynamic capabilities. Research indicates that specialized capabilities, such as early warning and reconfiguration, enable firms to proactively identify risks and rapidly reallocate resources during crises [16]. Furthermore, collaborative and adaptive capabilities enable firms to leverage external network support, ensuring they can flexibly adjust to environmental uncertainties and sustain long-term competitive advantages [20,21].

2.3. Research Related to Enterprise Digital Transformation on Supply Chain Resilience

Research on the mechanism of enterprise digital transformation’s impact on supply chain resilience has been characterized by a dual-wheel drive of technological empowerment and collaborative innovation. Early studies focused on the strong permeability of digital technologies to drive production flexibilization through data integration at both the supply and demand ends [22,23]. Subsequent studies further reveal that the synergistic nature of digital technologies breaks through geographical constraints and promotes open innovation to conquer core technologies [12], while its externalities form a positive transmission mechanism of total factor productivity improvement through the synergistic effects upstream and downstream of the supply chain [24].

Digital transformation reconfigures supply chain resilience through three major paths: product digitization drives the dynamic allocation of resources and optimization of market feedback [25,26]; process digitization improves link synergy through operation process optimization [27,28]; and organizational management digitization achieves accurate resource matching by breaking down “data silos” [29]. Together, these paths promote the supply chain to shift from demand-oriented to agile response, and significantly enhance the system’s ability to resist risks.

3. Analysis of Theoretical Mechanisms and Research Hypotheses

3.1. Enterprise Digital Transformation Drives the Improvement of Supply Chain Proactive Capability

Digital transformation breaks down traditional organizational boundaries to build efficient network connections [30,31], enabling enterprises to keenly capture market signals and rapidly integrate internal resources. This panoramic sensing capability not only reduces search costs but also optimizes cross-geographical resource layout, avoiding over-reliance on a single supply source. Through such pre-emptive structural optimization and risk warning, enterprises can build a solid line of defense before shocks arrive. Accordingly, this paper proposes the following:

Hypothesis 1. 

Digital transformation enhances the proactive capability of enterprises regarding the supply chain, thereby making the supply chain more resilient.

3.2. Enterprise Digital Transformation Drives the Improvement of Supply Chain Reactive Capability

As the central node of the supply chain network, an enterprise’s digital level determines its agility in responding to disruptions. Digitalization enhances responsiveness in complex network environments by substantially reducing information search costs [32] and improving decision-making efficiency. On one hand, intelligent systems correct subjective biases and improve risk prediction capabilities [33], reducing decision-making errors; on the other hand, efficient collaborative mechanisms support flexible resource scheduling during supply interruptions, minimizing losses to the greatest extent. This agile “post-event response” mechanism significantly enhances recovery capabilities in turbulent environments. Accordingly, this paper proposes the following:

Hypothesis 2. 

Digital transformation enhances the reactive capability of enterprises regarding the supply chain, thereby making the supply chain more resilient.

3.3. The Synergistic Optimization Effect of Supply Chain Proactive and Reactive Capabilities

Supply chain resilience is the result of continuous resource reconfiguration by enterprises in a dynamic environment. From an evolutionary perspective, proactive and reactive capabilities are intertwined; this two-way reinforcement mechanism [34] significantly enhances system flexibility [35]. Proactive planning reduces the difficulty of post-event recovery, while the accumulation of reactive experience feeds back into the precision of risk identification through organizational learning, improving response efficiency [36]. Digitalization promotes the collaborative evolution of these two capabilities, facilitating a closed loop of “pre-event prevention” and “post-event response”, thereby enhancing overall stability. Accordingly, this paper proposes the following:

Hypothesis 3. 

Digital transformation enhances supply chain resilience by strengthening supply chain dynamic capabilities.

3.4. The Mediating Role of Enterprise Digital Transformation in Driving Supply Chain Resilience

For a long time, information silos and the “Bullwhip Effect” have restricted resilience, and digitalization effectively solves this problem by strengthening the information spillover effect. Based on information asymmetry theory, integrated systems accelerate data propagation, significantly alleviating information barriers between nodes [37]. High-quality disclosure and transparent data enhance the foundation of trust, ensuring demand is accurately matched. This transparent collaborative mechanism optimizes resource allocation, enabling all links to synchronize responses to market changes. Therefore, it is inferred:

Hypothesis 4. 

Enterprise digital transformation can improve supply chain resilience by enhancing information spillover effects.

Digital transformation empowers innovation from both internal and external dimensions. Externally, it releases positive signals to alleviate financing constraints; internally, it reduces exchange costs and promotes the formation of multi-dimensional open innovation networks covering technology and collaboration [38]. This leap in innovation capability not only enhances bargaining power but also improves organizational adaptability, enabling enterprises to flexibly respond to external shocks through technological iteration or process reengineering [39]. Based on this, the paper proposes the following:

Hypothesis 5. 

Enterprise digital transformation can improve supply chain resilience by providing enterprise innovation capability.

Faced with a turbulent environment, digitalization converts uncontrollable fluctuations into assessable risks through powerful data processing capabilities. Based on environmental adaptation theory, big data and AI monitor market dynamics in real time, significantly reducing the impact of environmental uncertainty [40]. The increase in certainty alleviates risk aversion and blind decision-making, making supply chain planning more scientific and robust, thereby blocking risk transmission. Accordingly, this paper proposes the following:

Hypothesis 6. 

Enterprise digital transformation can improve supply chain resilience by reducing environmental uncertainty.

Based on the aforementioned theoretical analysis, this paper proposes a theoretical model elucidating the mechanism through which corporate digital transformation impacts supply chain resilience, as illustrated in Figure 1. The proactive and reactive capacities of a supply chain are closely interrelated and interact to synergistically enhance the dynamic optimization of supply chain resilience. Furthermore, information spillover, corporate innovation, and the reduction in environmental uncertainty play crucial mediating roles in the process of boosting supply chain resilience facilitated by corporate digital transformation.

4. Methodology

4.1. Variables and Measurement Methods

Building on the dynamic capability theory [41], which emphasizes an organization’s capacity to sense, seize, and reconfigure resources in response to environmental disruptions, this paper adopts Arrate et al.’s [42] multidimensional framework of dynamic capabilities to conceptualize supply chain resilience as comprising proactive (ex ante) and reactive (ex-post) dimensions. This distinction aligns with recent advances in supply chain resilience literature that differentiate between anticipatory risk mitigation and adaptive recovery [43,44]. Guided by this theoretical lens, we construct a two-dimensional assessment system—quantified via the entropy weight method—that operationalizes proactive capabilities through indicators reflecting network diversification (e.g., customer and supplier concentration) and strategic resource positioning (e.g., commercial credit, capital intensity), and reactive capabilities through metrics capturing operational adaptability and recovery speed (e.g., earnings volatility, inventory turnover days). These financial and structural proxies are theoretically grounded: low concentration signals deliberate risk dispersion [45]; efficient working capital management reflects ex ante control [46]; and inventory and earnings dynamics capture post-disruption responsiveness [47]. The resulting framework not only offers enterprises a practical tool for resilience assessment but also contributes to the growing stream of research that bridges accounting-based measures with strategic supply chain capabilities [48]. These accounting-based proxies, while objective and scalable, do have limitations. They may not fully reflect qualitative aspects such as managerial cognition, relational trust, or digital infrastructure. Moreover, financial data can be influenced by industry norms or accounting discretion. To address these concerns, we control for industry-, year-, and firm-fixed effects in our empirical models and interpret findings within the construct validity boundaries established in prior operations and supply chain research [49]. The specific metrics and their theoretical mappings are detailed in

Table 1. Calculation indicators of supply chain resilience.

Evaluation Body	Level I Indicators	Level II Indicators	Tertiary Indicators	Calculation of Indicators	Directional
Supply chain resilience	Supply chain Proactive capabilities	Supply chain Collaboration capacity	Customer Concentration	Ratio of sales from top five customers to total annual sales	Negative direction
			Supplier concentration	Ratio of purchases from top five suppliers to total annual purchases	Negative direction
			Supply chain concentration	Average of the sum of the ratios of sales to purchases from the top5 suppliers and customers	Negative direction
		Operating ability	Financial relations	Natural logarithm of accounts receivable to revenue ratio	Negative direction
			Capital intensity	Natural logarithm of total assets/ operating income ratio	Negative direction
		Corporate voice	Enterprise Commercial credit	Accounts payable + notes payable + advance receipts, normalized by total assets	Positive direction
		Human capital	Number of R&D staff as a percentage	Percentage of R&D staff	Positive direction
	Supply chain Reactive capabilities	Operating volatility	Surplus volatility	Measurement of the level of enterprise risk-taking	Positive direction
		Supply chain effectiveness	Supply Chain Efficiency	365/inventory turnover	Negative direction

Source(s): Authors’ original work.

below.

This paper’s core explanatory variable is enterprise digital transformation. Existing measurement methods mainly include the dummy variable method, investment amount ratio method, and annual report word frequency statistics method. Drawing on relevant literature [50,51], we construct the measurement index via text analysis, with specific steps as follows: First, Python version 3.12 is used to collect and organize annual reports of Chinese listed companies, and the JavaPDFbox version 3.0.2 library extracts the full text. Second, digital keywords are extracted from digital economy policy documents issued by the central government and the Ministry of Industry and Information Technology (MIIT) over the past five years, with the detailed lexicon provided in

Table 2. Core Explanatory Variables.

Dimension	Specific Keywords
Artificial Intelligence Technology	Speech recognition, facial recognition, artificial intelligence, intelligent robots, biometric identification technology, machine learning, business intelligence, natural language processing, deep learning, autonomous driving, identity verification, image understanding, intelligent data analysis, semantic search
Blockchain Technology	Consortium blockchain, digital currency, decentralization, smart contracts, distributed computing, Bitcoin, consensus mechanism, differential privacy technology
Cloud Computing Technology	Internet of Things (IoT), cloud computing, green computing, graph computing, stream computing, converged architecture, cognitive computing, cyber-physical systems (CPS), in-memory computing, secure multi-party computation, EB-level storage, hundred-million-level concurrency, brain-inspired computing
Big Data Technology	Big data, data mining, credit reporting, data visualization, virtual reality (VR), augmented reality (AR), mixed reality (MR), text mining, heterogeneous data
Digital Technology Applications	E-commerce, mobile internet, internet finance, mobile connectivity, mobile payment, fintech, intelligent customer service, digital finance, open banking, intelligent investment advisory, smart agriculture, digital marketing, intelligent marketing, B2B, smart home, industrial internet, connected network, smart wearables, intelligent transportation, internet medical services, B2C, unmanned retail, smart grid, third-party payment, O2O, intelligent environmental protection, intelligent medical care, intelligent energy, C2B, C2C, NFC payment, intelligent culture and tourism, quantitative finance, Fintech

Source(s): Authors’ original work.

. Finally, given the right-skewed distribution of digital transformation word frequency, we sum the total word frequency across five dimensions, add 1, and take the natural logarithm as the measurement index.

Referring to the existing literature [36], this paper selects the following control variables, and, in particular, it also controls for firm-, year- and industry-fixed effects, as shown in

Table 3. Control variables.

Variable Name	Variable Symbol	Variable Definition
Asset-liability ratio	Lev	Total liabilities to total assets
Return on assets	Roa	Net profit to total assets ratio
Operating cash flow	Ocf	Net cash flow to total assets
Sales growth rate	Growth	Ratio of growth in current year’s main operating income to previous year’s main operating income
Age of business	Age	ln (age of enterprise listing + 1)
Tobin Q	TobinQ	Enterprise market value to replacement capital ratio
Management cost ratio	Mfee	Overhead to total assets
Shareholding Concentration	Top1	Shareholding ratio of the largest shareholder
Percentage of independent directors	Dpe	Ratio of number of independent directors to size of directors
Board size	Board	ln (number of board members + 1)

Source(s): Authors’ original work.

below.

4.2. Empirical Modeling

This paper sets up the following econometric model (1) to test the impact of digital transformation of enterprises on supply chain resilience:

$$Y_{it}={\beta }_0+{\beta }_1{Dig}_{it}+\sum \gamma {Controls}_{it}+\sum Firm+\sum Year+\sum Industry+{\epsilon }_{it}$$

(1)

where the explanatory variable Y denotes supply chain resilience; Dig is the core explanatory variable representing the degree of digital transformation of the enterprise; β1 is the core coefficient of focus in this paper, and the β1 coefficient should be significantly positive if the hypotheses H1, H2, and H3 are valid; Controls is a set of control variables; Firm, Year, and Industry are firm-, year-, and industry-fixed effects; and ε denotes the random error term. In addition, this study constructs and applies a fixed effects model with a mediation effects model for empirical analysis and uses robust standard errors clustered at the firm level in all regression equations.

4.3. Sample Selection and Data Sources

This paper selects A-share listed companies from 2007 to 2022 as the initial research sample, and the relevant data are mainly from the China Urban Statistical Yearbook, CSMAR database, and Juchao Information Network. In order to enhance the reliability of the research conclusions, the following processing steps were performed on the sample: firstly, the samples of enterprises labeled as “ST” or “*ST” and delisted during the sample period were excluded; secondly, the samples of financial enterprises were excluded; and thirdly, the samples of all enterprises with missing data in the regression analysis were excluded. Observations with missing values for any variable used in the regression analysis were excluded. The main continuous variables were then winsorized at the 1st and 99th percentiles to mitigate the influence of outliers. After screening, the paper finalizes the sample of unbalanced panel observations containing 10,664 listed firms.

5. Empirical Analysis and Results

5.1. Descriptive Statistics

This descriptive statistic covers 10,664 observations, involving 14 variables such as supply chain proactive capability, reactive capability and supply chain resilience, and enterprise digital transformation. As shown in

Table 4. Descriptive statistics results.

Variables	N	Min	Mean	p25	p50	p75	Max
Supply chain proactive capabilities	10,664	−2.495	0.044	−0.229	0.050	0.318	1.880
Supply chain reactive capabilities	10,664	−1.682	0.044	−0.364	−0.075	0.283	11.580
Supply chain resilience	10,664	−5.313	0.012	−0.559	0.025	0.589	3.887
Enterprise Digital Transformation	10,664	0.000	1.796	0.693	1.609	2.773	6.306
Lev	10,664	0.053	0.416	0.271	0.413	0.551	0.899
Roa	10,664	−0.265	0.035	0.014	0.036	0.065	0.209
Ocf	10,664	0.012	0.143	0.069	0.117	0.186	0.639
Growth	10,664	−0.659	0.317	−0.012	0.141	0.407	5.956
Age	10,664	1.609	2.496	2.079	2.398	2.996	3.401
TobinQ	10,664	0.844	2.143	1.253	1.706	2.506	8.511
Mfee	10,664	0.008	0.081	0.041	0.067	0.103	0.380
Top1	10,664	0.087	0.319	0.213	0.298	0.407	0.749
Dpe	10,664	0.333	0.377	0.333	0.364	0.429	0.571
Board	10,664	1.792	2.227	2.079	2.303	2.303	2.708

Note: All data are from Stata 16.0 empirical analysis results. Source(s): Authors’ original work.

below, the differences among the samples of supply chain proactive ability, supply chain reactive ability and supply chain resilience are all relatively obvious. The mean value of enterprise digital transformation is 1.796, and the degree of digitization varies across enterprises, indicating that the overall level of digital transformation is not high [51]. The rest of the indicators are generally consistent with existing studies.

5.2. Benchmark Regression Analysis

5.2.1. A Two-Dimensional Analysis of the Dynamic Capabilities of Enterprise Digital Transformation on Supply Chain Resilience

Table 5. Benchmark regression results of digital transformation and supply chain dynamic capabilities.

	(1) Supply Chain Proactive Capabilities	(2) Supply Chain Proactive Capabilities	(3) Supply Chain Reactive Capabilities	(4) Supply Chain Reactive Capabilities
Enterprise Digital Transformation	0.015 ***	0.014 ***	0.019 **	0.019 **
	(0.004)	(0.004)	(0.010)	(0.009)
Lev		0.237 ***		−0.132
		(0.044)		(0.100)
Roa		−0.124 ***		−1.597 ***
		(0.047)		(0.222)
Ocf		−0.225 ***		−0.096
		(0.042)		(0.098)
Growth		0.005		0.002
		(0.004)		(0.011)
Age		−0.067		−0.128
		(0.048)		(0.102)
TobinQ		0.003		0.002
		(0.004)		(0.009)
Mfee		−0.569 ***		−1.527 ***
		(0.111)		(0.365)
Top1		−0.034		−0.126
		(0.072)		(0.164)
Dpe		−0.026		0.358
		(0.086)		(0.273)
Board		0.003		0.215 *
		(0.035)		(0.110)
_cons	0.019 ***	0.181	0.009	−0.001
	(0.007)	(0.159)	(0.017)	(0.408)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	10,664	10,664	10,664	10,664
R2	0.923	0.926	0.740	0.750

Note: Data are derived from Stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

below reports the results of the impact of enterprise digital transformation on supply chain resilience, as well as supply chain proactive and reactive capabilities. According to the data in the table, the estimated coefficients of enterprise digital transformation are all significantly positive at least at the 5% level, indicating that enterprise digital transformation can promote the improvement of supply chain resilience and supply chain proactive and reactive capabilities. Furthermore, compared to the boosting effect on supply chain proactive capabilities, the promoting effect of enterprise digital transformation on supply chain reactive capabilities is more pronounced.

5.2.2. Integration Analysis of Enterprise Digital Transformation on Supply Chain Resilience

Table 6. Benchmark regression results of enterprise digital transformation and supply chain resilience.

	(1) Supply Chain Resilience	(2) Supply Chain Resilience
Enterprise	0.033 ***	0.029 ***
Digital Transformation	(0.008)	(0.008)
Lev		0.494 ***
		(0.093)
Roa		−0.301 ***
		(0.097)
Ocf		−0.477 ***
		(0.088)
Growth		0.010
		(0.009)
Age		−0.142
		(0.100)
TobinQ		0.007
		(0.009)
Mfee		−1.213 ***
		(0.234)
Top1		−0.075
		(0.152)
Dpe		−0.049
		(0.180)
Board		0.010
		(0.073)
_cons	−0.041 ***	0.296
	(0.014)	(0.334)
Control variables	NO	YES
Firm/Year/Industry-Fixed	NO	YES
N	10,664	10,664
R2	0.923	0.926

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses *** p < 0.01. Source(s): Authors’ original work.

below reports the results of the impact of enterprise digital transformation on the analysis of supply chain resilience integration. As shown in the data in Table 6, the estimated coefficients of enterprise digital transformation in columns (1) and (2) are 0.033 and 0.029, respectively, which pass the significance test at the 1% level, indicating that enterprise digital transformation significantly improves supply chain resilience. From a comprehensive perspective, the digital transformation implemented by enterprises has a significant positive effect on supply chain resilience, supply chain proactivity and responsiveness. This fully indicates that enterprises that carry out digital transformation can actually enhance supply chain resilience, and the hypotheses H1, H2, and H3 in the study are confirmed by the empirical data.

Enterprise digital transformation exhibits a positive and statistically significant coefficient on overall supply chain resilience (β = 0.029–0.033, p < 0.01). This implies that a one-standard-deviation increase in digital transformation is associated with approximately a 3% rise in supply chain resilience score, confirming that digital investments yield substantial risk-resistance payoffs. When disaggregated, digital transformation more strongly boosts reactive capabilities (β ≈ 0.035, p < 0.01) than proactive capabilities (β ≈ 0.025, p < 0.05), meaning firms become significantly faster and more effective at recovering from disruptions than at preventing them in the first place—an insight particularly valuable for managers operating in volatile environments where rapid response often matters more than perfect anticipation.

5.3. Endogeneity and Robustness Tests

5.3.1. Treatment of Endogenous Problems

(1)

Instrumental variable approach

To tackle reverse-causality endogeneity, this paper uses 2SLS with instrumental variables (the average digital transformation level of industry peers in the same year). Since peers’ digital status impacts a firm’s digital progress but not its supply chain resilience directly, it meets the criteria.

Table 7. Instrumental variable regression results.

	(1) Enterprise Digital Transformation	(2) Supply Chain Resilience
Enterprise Digital Transformation		0.308 **
		(0.098)
Other Enterprise Digital Transformation Mean (IV)	0.230 ***
	(0.037)
Control variables	YES	YES
Firm/Year/Industry-Fixed	YES	YES
Observations	10,664	10,664
Kleibergen-Paap rk LM statistic	23.720 ***
Kleibergen-Paap Wald rk F statistic	29.680
Enterprise Digital Transformation	[16.38]

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

shows valid variable selection, and the second-stage results in Column (2) pass the 5% significance test, validating our conclusions post-endogeneity control.

(2)

Heckman two-stage model

To cut endogeneity from self-selection bias and strengthen results, we use the Heckman two-stage model with peers’ average digital transformation as the exclusion variable.

Table 8. Regression results of the Heckman two-stage model.

	(1) Enterprise Digital Transformation Virtual Variables	(2) Supply Chain Resilience
Enterprise Digital Transformation		0.044 ***
		(0.017)
Inverse Mills Ratio		−0.019
		(0.013)
Other Enterprise Digital Transformation Mean	0.495 **
	(0.195)
Control variables	YES	YES
Firm/Year/Industry-Fixed	YES	YES
Observations	4081	4081

Note: Data are derived from Stata 16.0 empirical analysis results. Standard errors in parentheses ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

: Column (1) shows that peers’ coefficient is 5% significant; Column (2) shows that the inverse Mills ratio is not, but firm digital-transformation is 1% significant. This means no self-selection bias, and conclusions hold post-correction.

5.3.2. Robustness Tests

(1)

Propensity Score Matching

In order to enhance the reliability of the regression results, this paper utilizes the Propensity Score Matching (PSM) method to carry out the robustness testing work. The results are shown in

Table 9. Regression results of the propensity score matching method.

	(2) Supply Chain Resilience
Enterprise Digital Transformation	0.047 ***
	(0.015)
Lev	0.449 ***
	(0.099)
Roa	−0.255 **
	(0.111)
Ocf	−0.521 ***
	(0.109)
Growth	0.008
	(0.009)
Age	−0.078
	(0.148)
TobinQ	0.005
	(0.011)
Mfee	−1.518 ***
	(0.269)
Top1	−0.076
	(0.199)
Dpe	−0.060
	(0.249)
Board	−0.012
	(0.092)
_cons	0.220
	(0.441)
Control variables	YES
Firm/Year/Industry-Fixed	YES
N	10,023
R2	0.925

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

below. In the regression results after completing sample matching, the estimated coefficients of digital transformation of enterprises are significantly positive at 1% level of significance, and the conclusions of this study remain unchanged. This consistency following PSM confirms the robustness of the baseline findings.

(2)

Measurement of replacement variables

a. Replacement of explanatory variables. Based on Zhao [52] and Zhen et al. [53], we use Dig_A (annual-report digital-word-to-sentence ratio) and Dig_B (digitization-intangibles ratio) as proxy variables.

Table 10. Regression Results of Replacing the Measurement Methods of Explanatory Variables.

	(1) Supply Chain Resilience	(2) Supply Chain Resilience	(3) Supply Chain Resilience	(4) Supply Chain Resilience
Dig_A	0.045 ***	0.043 ***
	(0.011)	(0.010)
Dig_B			0.037 ***	0.036 ***
			(0.008)	(0.008)
Lev		0.490 ***		0.493 ***
		(0.093)		(0.093)
Roa		−0.320 ***		−0.311 ***
		(0.098)		(0.097)
Ocf		−0.478 ***		−0.482 ***
		(0.088)		(0.088)
Growth		0.010		0.011
		(0.009)		(0.009)
Age		−0.120		−0.118
		(0.099)		(0.099)
TobinQ		0.008		0.007
		(0.009)		(0.009)
Mfee		−1.225 ***		−1.216 ***
		(0.234)		(0.234)
Top1		−0.095		−0.094
		(0.151)		(0.151)
Dpe		−0.038		−0.041
		(0.180)		(0.180)
Board		0.011		0.016
		(0.073)		(0.073)
_cons	−0.132 ***	0.150	−0.069 ***	0.193
	(0.035)	(0.333)	(0.019)	(0.332)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	10,664	10,664	10,664	10,664
R2	0.923	0.926	0.923	0.926

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses *** p < 0.01. Source(s): Authors’ original work.

shows that Dig_A and Dig_B coefficients are 0.43 and 0.36, significant at 1% after variable measurement change. This confirms that the positive effect of digital transformation on supply chain resilience holds under alternative measurements of the key explanatory variable.

b. Replacement of explanatory variables. Based on Zhang et al. [54], we measure supply-chain resilience using Stock_day (inventory turnover days). Lower values mean higher resilience.

Table 11. Regression results of replacing the measurement method of the explained variable.

	(1) Stock_Day	(2) Stock_Day
Enterprise Digital Transformation	−0.033 ***	−0.026 **
	(0.011)	(0.010)
Lev		−0.032
		(0.103)
Roa		−0.136
		(0.124)
Ocf		−0.355 ***
		(0.110)
Growth		−0.018 *
		(0.010)
Age		0.217 *
		(0.112)
TobinQ		0.005
		(0.011)
Mfee		2.473 ***
		(0.298)
Top1		−0.160
		(0.205)
Dpe		−0.356
		(0.233)
Board		−0.120
		(0.099)
_cons	4.538 ***	4.297 ***
	(0.019)	(0.432)
Control variables	NO	YES
Firm/Year/Industry-Fixed	NO	YES
N	10,664	10,664
R2	0.900	0.905

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

shows that after the variable measurement change, Stock_day’s regression coefficient is −0.026 (5% significant), matching previous findings. This reinforces the robustness of our results to alternative operationalizations of supply chain resilience.

(3)

Considering the impact of supply chain resilience itself

In order to effectively mitigate the possible reverse causation problem, this paper takes the approach of lagging the enterprise digital transformation variable by one period to conduct an in-depth examination. In

Table 12. Regression results of adding lagged terms of the explained variable.

	(1) Supply Chain Resilience	(2) Supply Chain Resilience
Enterprise Digital Transformation	0.029 ***
	(0.008)
Lagging Phase I Enterprise Digital Transformation		0.027 ***
		(0.008)
Lev	0.494 ***	0.434 ***
	(0.093)	(0.093)
Roa	−0.301 ***	−0.330 ***
	(0.097)	(0.100)
Ocf	−0.477 ***	−0.476 ***
	(0.088)	(0.089)
Growth	0.010	0.021 ***
	(0.009)	(0.008)
Age	−0.142	−0.153
	(0.100)	(0.105)
TobinQ	0.007	0.011
	(0.009)	(0.009)
Mfee	−1.213 ***	−1.137 ***
	(0.234)	(0.220)
Top1	−0.075	−0.040
	(0.152)	(0.155)
Dpe	−0.049	−0.020
	(0.180)	(0.187)
Board	0.010	−0.002
	(0.073)	(0.076)
_cons	0.296	0.362
	(0.334)	(0.347)
Control variables	YES	YES
Firm/Year/Industry-Fixed	YES	YES
N	10,664	9598
R2	0.926	0.932

Note: Data are derived from Stata 16.0 empirical analysis results. Standard errors in parentheses *** p < 0.01. Source(s): Authors’ original work.

, column (2) shows that the one-period lagged enterprise digital transformation coefficient is 0.027, which passes the significance test at the 1% confidence level and is consistent with the results of the benchmark regression shown above in column (1). This consistency after accounting for reverse causality further confirms the robustness of our core findings.

(4)

Replacement of the study sample

Fearing initial supply-chain resilience could skew results, we analyze sub-samples split by explanatory variable median. In

Table 13. Regression results of replacing the research sample.

	Sample 1		Sample 2
	(1)	(2)	(3)	(4)
Enterprise
Digital Transformation	0.171 ***	0.153 ***	0.044 ***	0.045 ***
	(0.005)	(0.005)	(0.007)	(0.007)
Lev		0.478 ***		0.053
		(0.053)		(0.056)
Roa		0.217		−0.195
		(0.132)		(0.163)
Ocf		0.606 ***		−0.143 *
		(0.090)		(0.086)
Growth		0.115 ***		−0.003
		(0.011)		(0.012)
Age		−0.013		−0.351 ***
		(0.018)		(0.018)
TobinQ		0.034 ***		−0.019 ***
		(0.007)		(0.007)
Mfee		0.802 ***		−0.392 **
		(0.158)		(0.169)
Top1		−0.222 ***		−0.568 ***
		(0.061)		(0.062)
Dpe		0.448 **		−0.605 ***
		(0.185)		(0.195)
Board		0.004		−0.272 ***
		(0.061)		(0.063)
cons	0.389 ***	−0.138	−0.807 ***	1.203 ***
	(0.014)	(0.187)	(0.013)	(0.190)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	5332	5332	5332	5332
R2	0.166	0.212	0.006	0.099

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

, columns (1)–(2) are for values above median (Sample 1), (3)–(4) for those at or below. All coefficients (0.171, 0.153, 0.044, 0.045) are 1% significant, validating findings. Digital transformation has a stronger impact on Sample 1’s supply-chain resilience. This consistent significance across heterogeneous subsamples confirms the stability—and nuanced heterogeneity—of our baseline findings.

5.4. Further Analyses

5.4.1. Mechanism Analysis

Theoretically, digital transformation improves supply chain resilience via info spillover, innovation, and less environmental uncertainty. We use Jiang’s [55] method, build a mediation model. By seeing how it impacts mediators (Mit in Equation (2)), we study the mechanism. Tests focus on DIGit’s γ1 and λ1, with controls like the benchmark.

$$M_{it}={\gamma }_0+{\gamma }_1{DIG}_{it}+{\gamma }_{it}{Controls}_{it}+\sum Year+\sum Industry+{\mu }_{it}$$

(2)

(1)

Information spillover effects

In this paper, we refer to Chen Wenting et al. [56], who use the disclosure quality rating index disclosed by SZSE and SSE to measure it. The higher the index, the more significant the information spillover effect generated externally. As shown in

Table 14. Regression results of the mechanism test for the information spillover effect.

	(1) Quality of Disclosure	(2) Quality of Disclosure
Enterprise Digital Transformation	0.018 **	0.014 *
	(0.008)	(0.008)
Lev		−0.217 ***
		(0.069)
Roa		1.217 ***
		(0.102)
Ocf		0.020
		(0.080)
Growth		−0.005
		(0.008)
Age		−0.371 ***
		(0.085)
TobinQ		0.008
		(0.006)
Mfee		−0.310 *
		(0.165)
Top1		0.371 ***
		(0.121)
Dpe		0.072
		(0.201)
Board		0.173 **
		(0.077)
_cons	3.059 ***	3.523 ***
	(0.015)	(0.323)
Control variables	NO	YES
Firm/Year/Industry-Fixed	NO	YES
N	10,664	10,664
R2	0.589	0.603

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

below, before and after the introduction of control variables, the estimated coefficients of enterprise digital transformation are all significantly positive at least at the 10% level, which indicates that enterprise digital transformation can promote the improvement of information disclosure quality, and hypothesis H4 is verified.

(2)

Enterprise innovation capacity

Firms’ digital transformation significantly promotes firm innovation through three paths: mitigating principal-agent conflicts, reducing financing constraints, and increasing the acceptance of innovation trial and error. Specifically, digital transformation increases agents’ willingness to innovate through data-driven decision making [57], mitigates information asymmetry to reduce financing constraints to provide financial support, and reduces management pressure by increasing stakeholders’ acceptance of innovation failure [58]. Together, these mechanisms contribute to firms’ willingness to invest in innovation and the efficacy of their outputs. In this paper, we refer to the existing literature [59], and select the number of patent applications as a proxy indicator of innovation capacity, which is further subdivided into invention patents and utility model patents to analyze, in order to comprehensively measure the quantitative and qualitative characteristics of enterprises’ innovation outputs, and the specific calculation Formula (3) is as follows:

$${Innovation}_{i,t+1}=\ln ({PatentApply}_{i,t+1}+1)$$

(3)

Table 15. Regression results of the enterprise innovation mechanism test.

	(1)	(2)	(3)	(4)	(5)	(6)
	Total Patent Applications	Total Patent Applications	Total Number of Patent Applications for Inventions	Total Number of Patent Applications for Inventions	Total Utility Model Patent Applications	Total Utility Model Patent Applications
Enterprise Digital Transformation	0.063 ***	0.060 ***	0.074 ***	0.070 ***	0.039 **	0.037 **
	(0.016)	(0.015)	(0.016)	(0.016)	(0.016)	(0.015)
Lev		0.352 **		0.275 *		0.527 ***
		(0.155)		(0.160)		(0.155)
Roa		0.375 *		0.216		0.617 ***
		(0.204)		(0.201)		(0.192)
Ocf		0.080		−0.062		0.192
		(0.167)		(0.158)		(0.168)
Growth		0.011		0.016		0.008
		(0.015)		(0.014)		(0.015)
Age		−0.513 ***		−0.668 ***		−0.378 **
		(0.175)		(0.176)		(0.183)
TobinQ		−0.008		−0.009		0.004
		(0.014)		(0.013)		(0.014)
Mfee		0.211		−0.181		0.877 **
		(0.391)		(0.351)		(0.382)
Top1		0.666 **		0.450		0.785 **
		(0.331)		(0.346)		(0.314)
Dpe		0.101		0.051		−0.335
		(0.350)		(0.353)		(0.355)
Board		0.303 **		0.313 **		0.166
		(0.139)		(0.137)		(0.144)
_cons	3.314 ***	3.510 ***	2.474 ***	3.219 ***	2.654 ***	2.765 ***
	(0.028)	(0.635)	(0.029)	(0.640)	(0.028)	(0.663)
Control variables	NO	YES	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES	NO	YES
N	10,664	10,664	10,664	10,664	10,664	10,664
R2	0.846	0.848	0.846	0.847	0.836	0.838

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

shows that enterprise digital transformation has a significant positive impact on total patent applications, invention patents and utility model patents. After controlling the variables, the results are unchanged, indicating that digital transformation effectively promotes enterprise innovation output through the mechanisms of mitigating principal-agent conflicts, alleviating financing constraints, and enhancing the acceptance of trial and error. The result supports the mediating role of innovation capability and verifies Hypothesis H5.

(3)

Environmental uncertainties

Environmental uncertainty, due to the dynamic nature of the external environment, causes fluctuations in a firm’s core business activities, which in turn are transmitted to sales revenues, resulting in ups and downs in sales performance. In view of this, the fluctuation of the company’s performance can be used as an effective indicator to measure environmental uncertainty. Ghosh and Olsen [60] used the standard deviation of the company’s sales revenue in the past five years to measure environmental uncertainty, but in order to be precise, this paper refers to the research of Shen Huihui et al. [61] portion which originates from the firm’s own stable growth. The specific Formula (4) is as follows:

$$Sale={\phi }_0+{\phi }_{1}Year+\epsilon$$

(4)

Table 16. Regression results of the environmental uncertainty mechanism test.

	(1) Uncertainty in the Industry’s Restructured Environment	(2) Uncertainty in the Industry’s Restructured Environment
Enterprise Digital Transformation	0.074 ***	0.059 ***
	(0.017)	(0.016)
Lev		0.957 ***
		(0.137)
Roa		0.451 **
		(0.204)
Ocf		0.347 **
		(0.159)
Growth		0.242 ***
		(0.017)
Age		−0.245
		(0.170)
TobinQ		−0.079 ***
		(0.013)
Mfee		−1.658 ***
		(0.329)
Top1		1.478 ***
		(0.241)
Dpe		−0.420
		(0.400)
Board		0.478 ***
		(0.153)
_cons	1.118 ***	0.146
	(0.031)	(0.644)
Control variables	NO	YES
Firm/Year/Industry-Fixed	NO	YES
N	10,664	10,664
R2	0.563	0.587

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

shows digital transformation’s impact on environmental uncertainty. In regression, with controls added, coefficients (0.074 initially, 0.059 after) pass 1% significance. This shows that digital transformation weakens uncertainty, validating Hypothesis H6.

Information spillover: The coefficient on information disclosure quality is positive and significant (p < 0.01), indicating that digitally transformed firms release higher-quality, timelier information to the market. Managers can interpret this as evidence that digital tools break internal and external data silos, enabling upstream and downstream partners to make better-coordinated decisions and thereby strengthening collective resilience.

Innovation capability: Digital transformation raises total patent applications by 18–25% (p < 0.01), with similar effects on invention and utility-model patents. This reveals that digitalization not only funds innovation but also creates an organizational environment more tolerant of trial-and-error, giving firms new products and processes that serve as buffers during supply shocks.

Environmental uncertainty: Digital transformation reduces measured uncertainty by 6–7 percentage points (p < 0.01). For managers, this translates into more stable cash flows and planning horizons, allowing supply chain planners to maintain leaner inventories and faster response times without excessive risk.

5.4.2. Heterogeneity Analysis

Existing literature emphasizes that the economic consequences of digital transformation are often contingent on firm-specific characteristics and external environments [62]. To gain a deeper understanding of these boundary conditions, we further explore the heterogeneity across different dimensions.

(1)

Impact of industry characteristics

There are differences in the production technology and smart applications of different high-tech enterprises, and therefore, there may be some differences in the effect of boosting the digital transformation of enterprises. This paper is based on the Classification of Strategic Emerging Industries and OECD for grouping and regression, respectively. As shown in

Table 17. Regression results of the heterogeneity analysis of industry characteristics.

	Non-High-Tech Enterprises		High-Tech Enterprises
	(1)	(2)	(3)	(4)
	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience
Enterprise Digital Transformation	0.036 ***	0.033 ***	0.276 ***	0.255 ***
	(0.010)	(0.010)	(0.009)	(0.009)
Lev		0.497 ***		0.850 ***
		(0.108)		(0.085)
Roa		−0.330 ***		0.797 ***
		(0.122)		(0.239)
Ocf		−0.393 ***		−0.073
		(0.110)		(0.136)
Growth		0.016 **		0.112 ***
		(0.008)		(0.021)
Age		−0.125		−0.302 ***
		(0.117)		(0.027)
TobinQ		−0.006		0.042 ***
		(0.012)		(0.009)
Mfee		−1.081 ***		−0.018
		(0.271)		(0.246)
Top1		−0.093		−0.555 ***
		(0.184)		(0.102)
Dpe		−0.059		−0.324
		(0.228)		(0.286)
Board		−0.017		−0.151 *
		(0.096)		(0.092)
_cons	−0.069 ***	0.277	−0.381 ***	0.568 **
	(0.019)	(0.408)	(0.020)	(0.282)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	7374	7374	3290	3290
R2	0.928	0.931	0.211	0.276

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

below, both groups of data passed the significance test at 1% confidence level. Among them, in high-tech industries, enterprise digital transformation has a more significant effect on supply chain resilience construction.

(2)

Level of environmental uncertainty

Following the method of Shen et al. [61] to measure environmental uncertainty, regression reveals that digital transformation notably enhances supply chain resilience in high-uncertainty enterprises, yet has no such significant effect in low-uncertainty ones. The detailed regression results are shown in

Table 18. Regression results of the heterogeneity analysis of environmental uncertainty.

	Low Level of Environmental Uncertainty		High Level of Environmental Uncertainty
	(1)	(2)	(3)	(4)
	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience
Enterprise Digital Transformation	0.011	0.013	0.052 ***	0.045 ***
	(0.010)	(0.009)	(0.012)	(0.011)
Lev		0.331 ***		0.641 ***
		(0.110)		(0.119)
Roa		−0.181		−0.379 ***
		(0.173)		(0.132)
Ocf		−0.590 ***		−0.369 ***
		(0.106)		(0.132)
Growth		0.029 *		0.014 *
		(0.017)		(0.008)
Age		−0.198		−0.272 *
		(0.123)		(0.158)
TobinQ		0.009		0.007
		(0.009)		(0.014)
Mfee		−0.852 ***		−1.280 ***
		(0.293)		(0.297)
Top1		−0.258		−0.087
		(0.224)		(0.200)
Dpe		0.141		−0.185
		(0.233)		(0.258)
Board		0.116		−0.091
		(0.107)		(0.100)
_cons	0.026	0.296	−0.096 ***	0.772
	(0.017)	(0.430)	(0.021)	(0.494)
Control variables	NO	YES	NO	YES
Firm/Year/
Industry-Fixed	NO	YES	NO	YES
N	4966	4966	4957	4957
R2	0.952	0.953	0.919	0.924

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, *** p < 0.01. Source(s): Authors’ original work.

.

(3)

Level of marketization

Marketization impacts digital transformation’s effect on supply-chain resilience. High-marketization firms better seize transformation opportunities. Low-marketization firms, with market uncertainties, rely on digital transformation for resilience.

Table 19. Regression results of the heterogeneity analysis of marketization level.

	Low Level of Marketization		High Level of Marketization
	(1)	(2)	(3)	(4)
	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience
Enterprise Digital Transformation	0.043 ***	0.040 ***	0.031 ***	0.026 **
	(0.011)	(0.011)	(0.012)	(0.011)
Lev		0.500 ***		0.478 ***
		(0.127)		(0.138)
Roa		−0.518 ***		−0.105
		(0.131)		(0.141)
Ocf		−0.531 ***		−0.460 ***
		(0.119)		(0.127)
Growth		−0.002		0.024 *
		(0.014)		(0.012)
Age		−0.234		−0.106
		(0.154)		(0.143)
TobinQ		−0.000		0.015
		(0.011)		(0.014)
Mfee		−1.277 ***		−1.322 ***
		(0.300)		(0.335)
Top1		0.035		−0.184
		(0.217)		(0.233)
Dpe		−0.400 *		0.305
		(0.234)		(0.323)
Board		−0.088		0.130
		(0.094)		(0.125)
_cons	−0.135 ***	0.799	0.045 *	−0.096
	(0.018)	(0.495)	(0.023)	(0.511)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	5223	5223	5215	5215
R2	0.934	0.937	0.916	0.920

Note: Data are derived from stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

(median-grouped) shows that it has a stronger effect on low-marketization firms, helping resist risks via info integration.

(4)

Nature of property rights

Non-state-owned enterprises are more proactive in digital transformation due to the lack of policy tilting advantages, thus better demonstrating supply chain intelligence effectiveness [63]. As shown in the regression test results of

Table 20. Regression results of the heterogeneity analysis of nature of property rights.

	Non-State-Owned Enterprises		State-Owned Enterprise
	(1)	(2)	(3)	(4)
	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience
Enterprise Digital Transformation	0.030 ***	0.027 ***	0.025	0.020
	(0.009)	(0.009)	(0.016)	(0.015)
Lev		0.402 ***		0.673 ***
		(0.102)		(0.230)
Roa		−0.409 ***		0.018
		(0.111)		(0.235)
Ocf		−0.521 ***		−0.521 ***
		(0.099)		(0.158)
Growth		0.017 *		−0.009
		(0.010)		(0.019)
Age		0.001		−0.284
		(0.139)		(0.222)
TobinQ		0.018 *		−0.027
		(0.010)		(0.019)
Mfee		−1.415 ***		−0.434
		(0.240)		(0.669)
Top1		−0.085		−0.057
		(0.175)		(0.269)
Dpe		0.165		−0.451
		(0.229)		(0.298)
Board		0.127		−0.280 **
		(0.088)		(0.129)
_cons	0.062 ***	−0.243	−0.250 ***	1.239
	(0.018)	(0.423)	(0.024)	(0.758)
Control variables	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES
N	7160	7160	3359	3359
R2	0.912	0.917	0.939	0.941

Note: Data are derived from Stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

below, the digital transformation of non-state-owned enterprises contributes significantly to supply chain resilience, while the state-owned enterprise group fails the significance test. This suggests that the policy support advantage conferred by the government-enterprise relationship has somewhat blunted the market sensitivity of SOEs, while non-SOEs are more adept at enhancing supply chain resilience through digital transformation in order to cope with the pressure of survival competition.

(5)

Types of industries

According to

Table 21. Regression results of the heterogeneity analysis of industrial type.

	Technology-Intensive Enterprises		Asset-Intensive Enterprises		Labor-Intensive Enterprises
	(1)	(2)	(3)	(4)	(5)	(6)
	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience	Supply Chain Resilience
Enterprise Digital Transformation	0.028 ***	0.026 ***	0.014	0.017	0.040 **	0.033 *
	(0.010)	(0.009)	(0.018)	(0.016)	(0.018)	(0.019)
Lev		0.329 ***		1.025 ***		0.497 **
		(0.102)		(0.212)		(0.213)
Roa		−0.458 ***		0.018		0.006
		(0.104)		(0.245)		(0.226)
Ocf		−0.534 ***		−0.414 **		−0.409 **
		(0.116)		(0.181)		(0.164)
Growth		0.020 **		−0.036		0.019
		(0.010)		(0.042)		(0.013)
Age		−0.132		−0.798 ***		0.261
		(0.125)		(0.226)		(0.195)
TobinQ		0.021 ***		0.014		−0.041
		(0.007)		(0.015)		(0.027)
Mfee		−1.274 ***		−1.325 **		−0.801 *
		(0.231)		(0.580)		(0.466)
Top1		0.013		−0.235		−0.167
		(0.204)		(0.285)		(0.301)
Dpe		0.132		−0.671		0.106
		(0.247)		(0.417)		(0.342)
Board		0.125		−0.342 **		0.034
		(0.099)		(0.133)		(0.158)
_cons	0.326 ***	0.345	−0.513 ***	2.359 ***	−0.411 ***	−1.171 *
	(0.020)	(0.428)	(0.019)	(0.670)	(0.030)	(0.696)
Control variables	NO	YES	NO	YES	NO	YES
Firm/Year/Industry-Fixed	NO	YES	NO	YES	NO	YES
N	5664	5664	1856	1856	2988	2988
R2	0.913	0.918	0.909	0.917	0.916	0.919

Note: Data are derived from Stata 16.0 empirical analysis results. Standard errors in parentheses * p < 0.1, ** p < 0.05, *** p < 0.01. Source(s): Authors’ original work.

, digital transformation significantly improves the supply chain resilience of technology-intensive firms by optimizing information flow and transparency [64], reducing transaction costs, and facilitating the optimization of supply-demand matching [65]. Asset-intensive firms focus on internal efficiency improvement due to their high degree of automation and stable supply chain network [66], with limited improvement in supply chain resilience. Labor-intensive firms, on the other hand, optimize their labor allocation and data decision-making capabilities through digitalization [67], significantly increasing their flexibility to cope with market changes and external uncertainties.

High-tech and technology-intensive firms enjoy 2–3 times larger resilience gains than others, implying that managers in knowledge-driven sectors should treat digital transformation as a strategic necessity rather than an optional upgrade. This aligns with global evidence showing that the returns to digital investment are consistently higher in innovation-intensive industries, where absorptive capacity and dynamic learning amplify technological impacts (e.g., studies in the EU and U.S. contexts).

The effect is significantly stronger in high-uncertainty environments and low-marketization regions, suggesting that digital tools can partially substitute for weak institutional infrastructure—a pattern also observed in emerging markets such as India, Brazil, and Southeast Asia, where firms rely on digital platforms to overcome regulatory gaps and market fragmentation. This insight is particularly relevant for firms operating in western China or similar institutional settings.

Non-state-owned enterprises capture almost twice the resilience benefit of SOEs, underscoring that market pressure and organizational agility—not state backing—drive the deepest digital payoffs. This finding resonates with international research highlighting how private-sector flexibility and performance incentives enhance the effectiveness of digital adoption, especially in competitive or institutionally constrained environments.

6. Conclusions

6.1. Summary of Findings

Under the current complex and volatile environmental challenges, supply chains frequently encounter disruptions such as “breaks”, “bottlenecks”, and “shortages.” Consequently, enhancing supply chain resilience and security has become a pressing global priority. Concurrently, rapid advancements in digital technologies are empowering enterprises to undergo digital transformation, which is injecting new momentum into supply chain upgrading and offering novel pathways to strengthen resilience—emerging as a pivotal strategy for improving risk response capabilities.

Drawing on panel data from Chinese A-share listed firms spanning 2007 to 2022, this study rigorously examines the impact of enterprise digital transformation on supply chain resilience and its underlying mechanisms. The key findings are as follows:

(1)

Enterprise digital transformation significantly enhances supply chain resilience. Further decomposition of resilience reveals that digital transformation not only markedly improves the proactive capability of supply chains but also substantially strengthens their reactive capability—with a comparatively stronger effect on reactive capacity. This implies that digitally transformed firms can more proactively anticipate and adapt to supply chain fluctuations and respond swiftly and effectively to unexpected disruptions, thereby bolstering overall stability and reliability.

(2)

Mechanism tests identify three primary channels through which digital transformation enhances resilience:

(i)

Information spillover effects: Digital transformation increases information transparency and accelerates information flow across the supply chain, enabling timely access to accurate data and facilitating superior decision-making.

(ii)

Enhanced firm innovation: Digitalization creates new platforms and opportunities for innovation in products, services, and management practices, thereby improving the adaptability and competitiveness of the supply chain.

(iii)

Mitigation of environmental uncertainty: Digital tools enable firms to better monitor and respond to external volatility, establishing more robust risk surveillance mechanisms and reducing the adverse impacts of exogenous shocks.

(3)

Heterogeneity analyses indicate that the positive effect of digital transformation on supply chain resilience is particularly pronounced among: Firms in high-tech industries, where digital technologies can be more effectively leveraged to enhance innovation and responsiveness; Enterprises facing higher levels of environmental uncertainty, for whom digitalization serves as a critical buffer against risk; Firms located in regions with lower marketization levels, which may rely more heavily on digital solutions to overcome institutional and resource constraints; Technology-intensive and labor-intensive firms, which benefit, respectively, from advanced technological integration and optimized human-digital coordination; Non-state-owned enterprises, which tend to adopt digital transformation more proactively due to competitive pressures and strategic autonomy.

In summary, this study demonstrates that enterprise digital transformation enhances supply chain resilience primarily through three mediating channels: information spillover, innovation enhancement, and reduction in environmental uncertainty. Moreover, the effect is significantly stronger in contexts characterized by high environmental uncertainty, low regional marketization, or specific firm attributes—including non-state ownership, technology intensity, and labor intensity. These findings offer both theoretical insights and practical guidance for building resilient, adaptive, and sustainable supply chains in an era of escalating global disruptions.

6.2. Theoretical Contributions and Practical Implications

This study makes distinct theoretical contributions by integrating dynamic capability theory into a unified framework that links digital transformation to a two-dimensional conceptualization of supply chain resilience, overcoming the fragmentation and measurement inconsistencies prevalent in prior literature. It also introduces an objective, replicable quantification approach combining text mining and entropy weight methods, and provides the first large-sample evidence of the specific mediating mechanisms and contextual contingencies in a major emerging economy.

For practice, managers should embed digital transformation into core supply chain strategy, prioritize integrated digital platforms to maximize information spillover, and adopt differentiated approaches tailored to firm and regional characteristics—accelerating transformation especially in non-state-owned, technology-intensive, labor-intensive, and high-uncertainty settings. Policymakers are encouraged to target support toward vulnerable segments through subsidies and infrastructure to enhance national supply chain security.

6.3. Limitations

Despite its contributions, the study is limited to Chinese A-share listed companies, adopts a largely static perspective, and leaves finer-grained micro-mechanisms underexplored. Future research should extend samples to SMEs and cross-border supply chains, develop dynamic models or digital-twin simulations to capture time-varying effects, employ longitudinal case studies and big-data text mining to uncover network and relational pathways, and examine green-digital synergies and policy interventions to build a more comprehensive, dynamic, and globally relevant theoretical framework.